Issue building vision tower

[lcernyo]

I am facing the following error when loading the model through:

from cli import HuatuoChatbot
bot = HuatuoChatbot(huatuogpt_vision_model_path)

See error message below:

[bxdgfw]

the same problem, have you solved it?

[0xYYQ]

I took a look at cli.py, and it looks like it uses path names to select the model if 'huatuogpt-vision-7b' in d.lower(): print(f'loading from {self.model_dir}') from llava.model.language_model.llava_qwen2 import LlavaQwen2ForCausalLM. So change the name of the folder where you store your model to "huatuogpt-vision-7b"